From Steel to Silicon

In the autumn of 1945, on the gutted third floor of what used to be a department store in central Tokyo, a thirty-seven-year-old engineer named Masaru Ibuka was trying to make broken radios sing again.

The building was a husk. The Shirokiya store, on a corner in Nihonbashi that had once been a fashionable place to buy kimono and Western pencils, had been hollowed out by American B-29s. The roof leaked. Whole floors were charred. But the iron frame still stood, which was more than could be said of most of central Tokyo, where the great fire raid of March 9, 1945 had burned out almost sixteen square miles in a single night and left more than a million people without homes. Ibuka had rented space inside the wreck because rent in a wreck was cheap, because the wreck still had electricity, and because he had decided, somewhere between the announcement of surrender on August 15 and the first frosts of autumn, that he was going to build something new in Japan and he was going to build it now.

The product he could actually sell that winter was an adaptor. During the war Japan had banned shortwave reception in civilian sets, sending inspectors and “radio consultants” through repair shops to check that ordinary domestic receivers could pick up only the longwave broadcasts of the state network. With the war over, a country that had been told for years that it was winning suddenly wanted to know what the rest of the world was actually saying. Ibuka and a handful of former colleagues from his wartime instrument company built small converter boxes that plugged into a longwave receiver and gave it back its missing bands. They sold as fast as he could make them. On October 6, 1945, the Asahi Shimbun ran a short item about the little workshop in the burned-out store, and that, more than anything else, is what set the rest of the century in motion. It was read in Nagoya by a young naval lieutenant named Akio Morita.

Morita was twenty-four. He came from one of the oldest sake-brewing families in Japan, the Moritas of Kosugaya, on the Chita peninsula south of Nagoya, who had been making rice wine since 1665. He was the eldest son. By every rule of his upbringing he should have been at home that autumn, putting on the dark suit of the fifteenth-generation heir and learning how to manage the brewery accounts. Instead he had been trained as a physicist at Osaka Imperial University, commissioned into the Imperial Japanese Navy, and assigned to a research committee on infrared-guided weapons, the so-called Ke-Go program, which is where, in 1944, he had met an older civilian engineer with a shy stoop and a habit of muttering at circuit diagrams. That engineer was Ibuka. They had liked each other immediately. Now Morita read his old colleague’s name in the newspaper and wrote to him. Ibuka wrote back.

By the spring of 1946 they had decided to throw in together. On May 7, with about twenty employees and a stake of 190,000 yen, much of it raised by Morita’s father, who agreed to bankroll a son who would not come home and brew sake, they incorporated a company in the third-floor wreck and called it Tokyo Tsushin Kogyo Kabushiki Kaisha, the Tokyo Telecommunications Engineering Corporation. Ibuka wrote the founding prospectus himself. Its first stated purpose was not a product line or a market. It was, he wrote, “to establish an ideal factory that stresses a spirit of freedom and open-mindedness, and where engineers with sincere motivation can exercise their technological skills to the highest level.” He was, in his own quiet way, declaring that Japan was going to invent things again.

It was an absurd document for the moment. Tokyo in 1946 had so little food that Ibuka and his men sometimes left the city on rucksack runs to the countryside to barter for rice and potatoes. The Supreme Commander for the Allied Powers, an American general named Douglas MacArthur, had imposed strict controls on Japanese industry, particularly on anything that might be repurposed for war. There was almost no copper. There was almost no rubber. Steel was rationed. Electricity was unreliable. And the components Ibuka most wanted, vacuum tubes, the glass bottles full of glowing filaments that were the brain of every radio and amplifier of the era, were, in the words of one of his early engineers, “hard to find in postwar Japan.” When you found them you paid black-market prices. When you ran them, they got hot.

The first thing the new company tried to manufacture was rice cookers. This was less foolish than it sounds. Rice was the staple of Japanese life, and after the war, with military demand gone and the country awash in idle hydroelectric capacity from dams built for munitions plants, electricity was the one industrial input that was actually plentiful. Ibuka designed a wooden tub with aluminum electrodes laid into its base. As long as the rice and water were wet, current flowed between the electrodes through the slurry, the resistance generated heat, and the rice cooked. When the water boiled away, the circuit broke and the cooker switched itself off.

The cooker did not work. Or rather, it worked sometimes. Depending on the kind of rice, the temperature of the water, and how the electrodes had been pressed into the wood, the device would deliver a meal raw on top, scorched on the bottom, or pasted to the electrodes in a glaze. Ibuka’s team built a hundred. They sold almost none. Morita, who had taken charge of the books, wrote off the inventory.

Their next experiment was a heated cushion, an electric zabuton, the flat square pad that Japanese sit on at low tables. This one was even simpler: nichrome wire stitched between layers of fabric and stuffing, plug in the wall, and your floor became warm. The cushions sold. They sold so well that Ibuka, embarrassed by the lack of any actual electronics in the device, had them branded under a fictitious company name, the Ginza Heating Company, so that buyers would not associate the cheap pad with the serious engineering firm he was trying to build. Then customers began reporting that the pads scorched their futons. A few caught fire. Ibuka pulled them off the market. He was relieved.

The first product that Tokyo Tsushin Kogyo built that actually behaved like telecommunications equipment was a vacuum-tube voltmeter, a precise instrument that measured small electrical signals using a tube as an amplifier. They were what radio engineers used to align a receiver, telephone technicians used to balance a line. Japan had almost no measurement equipment in 1946 because most of it had been either bombed or appropriated by the occupation. The Ministry of Communications, which was trying to put Japan’s telephone network back together, ordered fifty units. NHK, the national broadcaster, came in next. By 1947, the company was selling three or four dozen voltmeters a month. They paid the rent.

What the voltmeters also did, slowly, was force Ibuka and Morita to confront the limits of the medium they had bet their company on. Every voltmeter had to have a vacuum tube in it. Every shortwave converter had to have at least one vacuum tube in it. Every amplifier they ever sold to a courtroom or a school had to have several. And vacuum tubes, even when you could find them, were a miserable foundation on which to build a consumer electronics company.

A vacuum tube was a small evacuated glass bulb with a heated filament inside it. The filament boiled electrons off a metal cathode, the electrons were drawn through the vacuum to a positive plate, and a small grid in between let a tiny varying voltage modulate the much larger current flowing through the tube. That was the magic. It was what made radios receive, what made amplifiers amplify, what made the new digital computers in America tick. It was also what made all of that equipment heavy, hot, fragile, slow to start, and prone to die at random. The filament had to glow at red heat to work, which is why every tube radio of the 1940s took a moment to warm up and put out a faint orange light when you peered in the back. That heat had to go somewhere. The bigger the system, the worse the problem.

The cleanest illustration was already running in a basement in Philadelphia. The Electronic Numerical Integrator and Computer, ENIAC, completed by the U.S. Army at the Moore School of Electrical Engineering in 1945 and 1946, was the first general-purpose electronic computer. To do its work, it used roughly seventeen thousand vacuum tubes. ENIAC weighed thirty tons. It filled a room the size of a small gymnasium. It drew about 150 kilowatts of power, of which roughly 80 went simply to keeping the tubes’ filaments glowing, and another 20 went to fans trying to remove the heat the tubes were dumping into the room. With that many tubes operating at once, the failure of any single bottle would crash a calculation, and the engineers running the machine kept it powered up around the clock, because a cold start put thermal stress on the filaments and burned more tubes than ordinary running did. Even so, ENIAC suffered a tube failure on average every couple of days, and locating the bad tube among the seventeen thousand could take fifteen minutes of careful tracing. The machine was the most powerful calculator on earth and it was also a furnace held together by replacement parts.

This was the technology Ibuka and Morita had to work with. Scaled down to a radio set or a rice cooker rather than a thirty-ton calculator, the same physics played out at a smaller and more frustrating size. Vacuum tubes drained batteries within hours. They wanted hundreds of volts to run, which meant any portable device needed a stack of dry cells the size of a brick or a bulky transformer plugged into the wall. They were made of glass, which broke when dropped. The filaments aged: a tube that worked fine on the first day might be dim and noisy by the second year and dead by the fifth. Anything you built around them inherited their weaknesses. A radio big enough to carry on your shoulder. A “portable” record player you needed both hands for. Everything serious in electronics, in the year Tokyo Tsushin Kogyo opened its doors, was tube electronics, and tube electronics had a ceiling. You could feel it whenever you picked up your equipment. You could hear it in the low hum of every loudspeaker.

Morita, who would spend the next forty years thinking about consumer products, came to the question almost intuitively. The thing ordinary people wanted, he believed, was not bigger and louder. It was smaller and ever-present. A radio you could put in a coat pocket. A recording machine a child could lift. He could not say in 1947 exactly what the technology to do that would look like, but he knew it was not what he was building. Whenever he looked at the back of one of his own voltmeters, he was looking at a dead end with a faint orange glow.

Ibuka was thinking on a longer arc. He had been a research engineer all his working life, and he understood that the problem of tube electronics was not going to be solved by better tubes. Engineers had been making vacuum tubes smaller and more rugged for two decades, and every gain was immediately offset by the next jump in what people wanted to do with them. The country that figured out how to replace the tube with anything that did not need a glowing filament and an evacuated bulb and a brick of dry cells would own the second half of the twentieth century. Japan was not going to be that country in 1947. Japan, in 1947, was a half-starved archipelago under foreign occupation, the third floor of a burned department store, twenty employees, and a few hundred dollars of capital. But Ibuka and Morita read journals, and the journals had begun to carry rumors out of America that something interesting was happening at the laboratories of the American telephone monopoly, in a small town in New Jersey.

On a long, low brick campus in Murray Hill, a chain-smoking theorist and two experimentalists working under him had been pushing on the strange behavior of certain crystalline solids, semiconductors, materials neither good conductors of electricity nor good insulators but something controllable in between. They had spent the war years on radar and the years just after on a problem nobody outside the laboratory thought urgent: whether you could persuade a piece of solid matter to do the thing a vacuum tube did, only without the vacuum, the glass, the filament, or the heat.

In December 1947, while Morita was checking inventory in Nihonbashi and Ibuka was hunched over a voltmeter chassis, the three men in New Jersey would call their bosses into a small room, place a strip of gold foil and a sliver of germanium on a bench, and watch a needle on an oscilloscope jump.